[PATCH] D74320: GlobalISel: Add combine to narrow G_LSHR

[Matt Arsenault via Phabricator via llvm-commits]

arsenm created this revision.
arsenm added reviewers: dsanders, aditya_nandakumar, aemerson, paquette.
Herald added subscribers: kerbowa, Petar.Avramovic, volkan, hiraditya, rovka, nhaehnle, wdng, jvesely.
Herald added a project: LLVM.

Produce an unmerge to a narrower type and introduce a narrower shift
if needed. I wasn't sure if there was a better way to parameterize the
target's preferred shift type for the GICombineRule, so manually call
the combine helper.


https://reviews.llvm.org/D74320

Files:
  llvm/include/llvm/CodeGen/GlobalISel/CombinerHelper.h
  llvm/lib/CodeGen/GlobalISel/CombinerHelper.cpp
  llvm/lib/Target/AMDGPU/AMDGPUPreLegalizerCombiner.cpp
  llvm/test/CodeGen/AMDGPU/GlobalISel/combine-lshr-narrow.mir

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D74320.243540.patch
Type: text/x-patch
Size: 11059 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20200210/722679de/attachment.bin>